Abstract:  Fruit and vegetable drying is an important part in the processing of agricultural products, and the construction of accurate drying kinetics models has become a key direction in the drying field. In this paper, the application status of artificial neural network in fruit and vegetable drying was reviewed, the existing problems were analyzed and the prospects was made. The scenes of the artificial neural network in the drying process were classified into four parts such as water content prediction, quality detection, process optimization and control system, the application types and development innovations of each part were summarized. Further comparison was made between traditional drying models and artificial neural network models. Finally, the application scenarios of hybrid neural networks were introduced.  It is found that the artificial neural networks  is more accurate than the traditional drying models, and the hybrid neural networks combined with expert systems, fuzzy logic and other theories can provide accurate predictions. As a novel and efficient modeling technology, it can be widely applied in the optimization, control, automation and other fields of fruit and vegetable processing. The most widely used among them is the GA-BP neural network combined with genetic algorithms, where BP is responsible for prediction and GA is responsible for optimization. In such algorithms, not only can the results be accurately predicted but also the process can be optimized. This model is more suitable for fruit and vegetable drying and has broader development space in the future, with the hope that these discussions and analyses have reference significance for the field of fruit and vegetable drying.

Key words:  fruit and vegetable drying, neural network, drying kinetics model, error back propagation algorithm, moisture content prediction

摘要： 果蔬干燥是农产品加工中的重要环节，构建精确的干燥动力学模型成为干燥领域的重点方向。综述人工神经网络在果蔬干燥过程中的应用现状、分析存在的问题和做出展望。针对神经网络在干燥过程中的各种场景分类为四个部分：含水率预测、品质检测、工艺优化和控制系统方面，总结各部分的应用类型及发展创新；再对比传统干燥模型和人工神经网络模型；最后介绍混合神经网络的应用场景。发现人工神经网络比传统干燥模型更精确，且混合神经网络结合专家系统、模糊逻辑等理论能够提供准确的预测，作为一种新颖高效的建模技术，可以广泛应用于果蔬加工的优化、控制、自动化等领域。其中应用最广泛的就是与遗传算法结合的GA-BP神经网络，BP负责预测、GA负责寻优，在这样的算法中不仅可以精确预测结果还可以优化工艺。这样的模型更适合果蔬干燥且在未来有更广阔的发展空间，以期这些探讨和分析对果蔬干燥领域具有参考意义。

关键词: 果蔬干燥, 神经网络, 干燥动力学模型, 误差反向传播算法, 含水率预测